Dashboards were a big step forward. They took numbers out of spreadsheets and put them in front of the people who make decisions. Most organizations now have more dashboards than anyone can keep track of.

And yet the most common experience with a dashboard is looking at it, seeing something odd, and having a question it can't answer.

What dashboards are good at

Dashboards answer questions someone anticipated when they built the dashboard: revenue by month, pipeline by stage, orders by region. For monitoring a known set of measures, they are hard to beat. They are fast, consistent and shared.

Where they stop

The trouble is the next question.

Revenue dipped in the west region. Which customers? Which products? Since when? Compared with what? Each follow-up is a new filter, a new chart or, more often, a request to an analyst. Multiply that across every team and you get the familiar pattern: more dashboards, more reports, more waiting.

Three things tend to happen:

  • Dashboard sprawl. Each new question produces a new dashboard that is used a few times and then forgotten.
  • Analyst queues. Skilled people spend their days answering one-off questions instead of doing deeper work.
  • Decisions on stale answers. By the time the answer arrives, the moment to act has passed.

Conversation as an interface

Conversational analytics lets people ask the next question directly, in plain language, and get an answer from the same governed data the dashboards use:

  • Which customers in the west region are buying less than last quarter?
  • Of those, which were top-20 accounts last year?
  • What products did they stop buying?

Each answer builds on the previous one. Nobody builds a dashboard for a question that will be asked once.

It won't replace dashboards

Conversation is a poor way to watch the same ten numbers every morning. Dashboards will stay the right tool for monitoring. Conversational analytics covers the long tail: the specific, time-sensitive questions that dashboards were never going to anticipate.

The likely future is both: dashboards to notice, conversation to understand.

What it takes to trust it

A chart on a dashboard went through review before it was published. A conversational answer is generated on demand, so it needs other guarantees:

  • It should show which data it came from, and the logic behind it for anyone who wants to check.
  • It should respect who is allowed to see what.
  • It should say when a question is ambiguous or can't be answered from the data, instead of guessing.

Where QuerySafe fits

QuerySafe grew out of years of analytics work at MetricVibes, where we saw the same pattern again and again: the data existed, the dashboards existed, but the answer to the next question still took days. QuerySafe Intelligence connects to your BigQuery data and answers those questions in plain language. For a longer introduction, read What Is Conversational Business Intelligence?